It has long been known that the resistance of resistive materials used in resistors changes with temperature. This characteristic has been denoted as the "temperature coefficient of resistance" (TCR) and is measured by determining the actual resistance change for each degree of temperature change. The TCR is typically given in parts per million variation per degree centigrade, or ppm/.degree. C.
To assure predictability of operation of electronic devices using resistors, and to increase the precision and reliability of such devices, there have been many attempts to reduce the TCR of resistors towards "absolute zero". There is currently no method of achieving this absolute zero. Currently, the TCR is considered to be substantially zero when it is within the range of plus or minus 0.5 ppm/.degree. C.
In the past, precision wirewound resistors have been made by a costly process selecting equal value resistors and selectively matching negative and positive TCR's (which result from manufacturing variations) from a large collection of resistors. While initial TCR's could be controlled closely, long term resistance drift occurred which was unique for each resistor and which resulted in the overall resistance drifting with time. Further, due to the bulk of the individual resistors, close thermal coupling was and is not possible so the apparent ratio TCR (to be hereinafter defined) also changed with temperature gradients present.
In the past also, a typical approach for thin film resistors has been to find low TCR materials. A resistor of this type is shown in the Burger et al., Patent, U.S. Pat. No. 4,464,646. The resistor is made of tantalum, tantalum nitride, or tantalum oxinitride and is described as having "essentially" zero TCR. In fact, these materials have TCR's which range from plus or minus 100 ppm/.degree. C. which are many orders of magnitude from absolute zero or even substantially zero.
While Burger teaches a thin film circuit for controlling the temperature coefficient of resistance, it is directed towards providing a resistor which will be temperature dependent so it can act as a temperature sensor or a compensator for other elements. It does not teach or suggest a compound resistor having even substantially zero TCR since it presupposes that materials exist capable of providing such a TCR for the applications in which the Burger invention is used. For example, Burger specifies essentially zero TCR Tantalum which actually has a TCR of 80 ppm/.degree. C. plus or minus 10%.
Other attempts have been made to control TCR by the amount of material contained in a resistor as shown in the Baxter Patent, U.S. Pat. No. 4,375,056, or by control of the configuration of the resistor as shown in the Dorfield Patent, U.S. Pat. No. 4,079,349. However, the TCR's using these methods are still several orders of magnitude away from absolute zero or even substantially zero.
A further method of control is by the manipulation of the processing of the resistor materials, the annealing of the materials with temperature, or otherwise controlling the manufacture of the material and its deposition on substrates (in the case of thin film resistors) while forming the resistor. However, such process variations are complex, expensive, and cannot predictively and repeatably achieve the desired results. Thus, any given resistor manufactured by using materials so fabricated may still suffer various temperature deficiencies which cannot thereafter be corrected. Therefore, there has been a long term need for a resistor which has absolute zero TCR or which can be adjusted to have a substantially zero TCR after its fabrication has been completed.
When considering groups of resistors, which are called resistor arrays, there has been a long felt need for a resistor in which the TCR can be adjusted after fabrication to match or compensate for the TCR in other resistors of the array.
In some resistor arrays, it is not essential that all the resistors have an absolute or substantially zero TCR. It is more important that as temperature changes, the TCR's track each other or change resistances in parallel with changes in temperature. This important characteristic is called ratio TCR and is often expressed as the difference in the TCR's of various resistors in the resistor array. Since it is a difference of TCR's, it is also measured in ppm/.degree. C.
In the prior art, there's been a long felt need to minimize the ratio TCR in resistor arrays, and particularly to minimize the ratio TCR to zero or substantially zero for precision devices which is below 0.5 ppm/.degree. C.
In the past, very high precision resistors were generally hermetically sealed to reduce shifting in TCR values or resistances due to humidity or other environmental effects. This hermetic sealing made it impossible to change either the TCR of the resistance after the resistor array was completely fabricated. In essence, any final "adjustment" to refine any value was not physically possible.